Flash Deblurring

Overview

In this project, we aim to remove motion blur in images captured under low light environments. We utilize additional flash images to aid the blur kernel estimation and sharp image reconstruction. One example is shown below.

Blurred Image
Flash Image
Deblurred Image

Abstract

Motion blur due to camera shake is a annoying problem in low-light photography. In this paper, we propose a novel method to recover a sharp image from a pair of motion blurred and flash images, consecutively captured using a hand-held camera. We first introduce a robust flash gradient constraint by exploring the correlation between a sharp image and its corresponding flash image. Then we formulate our flash deblurring as solving a maximum-a-posteriori problem under the flash gradient constraint. We solve the problem by performing kernel estimation and non-blind deconvolution iteratively, leading to an accurate blur kernel and a reconstructed image with fine image details. Experiments on both synthetic and real images show the superiority of our method compared with existing methods.

Paper

Robust Flash Deblurring
Shaojie Zhuo, Dong Guo and Terence Sim
IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2010.
High-res PDF (36.8M) | Low-res PDF (7.6M) | BibTeX | Image Data | Poster (10.1M)

BibTeX

@inproceedings{ZhuoEtal10,
  author = {Shaojie Zhuo, Dong Guo and Terence Sim},
  title = {Robust Flash Deblurring},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
  year = {2010}
}